Upper crustal evolution across the Juan de Fuca ridge flanks · 4] The first to make the...

Upper crustal evolution across the Juan de Fuca ridge flanks

Mladen R. NedimovicDepartment of Earth Sciences, Life Sciences Centre, Dalhousie University, Halifax, Nova Scotia B3H 4J1, Canada([email protected])

Lamont-Doherty Earth Observatory, Columbia University, Palisades, New York 10964, USA

Suzanne M. Carbotte and John B. DieboldLamont-Doherty Earth Observatory, Columbia University, Palisades, New York 10964, USA

Alistair J. HardingScripps Institution of Oceanography, University of California, San Diego, La Jolla, California 92093, USA

J. Pablo CanalesWoods Hole Oceanographic Institution, Woods Hole, Massachusetts 02543, USA

Graham M. KentScripps Institution of Oceanography, University of California, San Diego, La Jolla, California 92093, USA

[1] Recent P wave velocity compilations of the oceanic crust indicate that the velocity of the uppermostlayer 2A doubles or reaches �4.3 km/s found in mature crust in <10 Ma after crustal formation. Thisvelocity change is commonly attributed to precipitation of low-temperature alteration minerals within theextrusive rocks associated with ridge-flank hydrothermal circulation. Sediment blanketing, acting as athermal insulator, has been proposed to further accelerate layer 2A evolution by enhancing mineralprecipitation. We carried out 1-D traveltime modeling on common midpoint supergathers from our 2002Juan de Fuca ridge multichannel seismic data to determine upper crustal structure at �3 km intervals along300 km long transects crossing the Endeavor, Northern Symmetric, and Cleft ridge segments. Our resultsshow a regional correlation between upper crustal velocity and crustal age. The measured velocity increasewith crustal age is not uniform across the investigated ridge flanks. For the ridge flanks blanketed with asealing sedimentary cover, the velocity increase is double that observed on the sparsely anddiscontinuously sedimented flanks (�60% increase versus �28%) over the same crustal age range of5–9 Ma. Extrapolation of velocity-age gradients indicates that layer 2A velocity reaches 4.3 km/s by�8 Ma on the sediment blanketed flanks compared to �16 Ma on the flanks with thin and discontinuoussediment cover. The computed thickness gradients show that layer 2A does not thin and disappear in theJuan de Fuca region with increasing crustal age or sediment blanketing but persists as a relatively lowseismic velocity layer capping the deeper oceanic crust. However, layer 2A on the fully sedimented ridge-flank sections is on average thinner than on the sparsely and discontinuously sedimented flanks (330 ± 80versus 430 ± 80 m). The change in thickness occurs over a 10–20 km distance coincident with the onset ofsediment burial. Our results also suggest that propagator wakes can have atypical layer 2A thickness andvelocity. Impact of propagator wakes is evident in the chemical signature of the fluids sampled by ODPdrill holes along the east Endeavor transect, providing further indication that these crustal discontinuitiesmay be sites of localized fluid flow and alteration.

G3G3GeochemistryGeophysics

Geosystems

Published by AGU and the Geochemical Society

AN ELECTRONIC JOURNAL OF THE EARTH SCIENCES

GeochemistryGeophysics

Geosystems

Article

Volume 9, Number 9

30 September 2008

Q09006, doi:10.1029/2008GC002085

ISSN: 1525-2027

ClickHere

for

FullArticle

Copyright 2008 by the American Geophysical Union 1 of 23

http://dx.doi.org/10.1029/2008GC002085

Components: 13,053 words, 14 figures, 2 tables.

Keywords: upper crustal evolution; multichannel seismics; traveltime modeling; reflection imaging; Juan de Fuca ridge

flanks.

Index Terms: 3035 Marine Geology and Geophysics: Midocean ridge processes; 3025 Marine Geology and Geophysics:

Marine seismics (0935, 7294); 7220 Seismology: Oceanic crust.

Received 2 May 2008; Revised 20 June 2008; Accepted 29 July 2008; Published 30 September 2008.

Nedimovic, M. R., S. M. Carbotte, J. B. Diebold, A. J. Harding, J. P. Canales, and G. M. Kent (2008), Upper crustal evolution

across the Juan de Fuca ridge flanks, Geochem. Geophys. Geosyst., 9, Q09006, doi:10.1029/2008GC002085.

1. Introduction

[2] The Earth’s oceanic crust crystallizes frommagmatic systems generated at mid-ocean ridges.For ridges with fast to intermediate spreading rates,the lower section of the oceanic crust is composedof layered and massive gabbros on top of which liediabase sheeted dykes and basaltic lavas of theupper oceanic crust. Tens of millions of years canpass until the oceanic crust formed at mid-oceanridges is subducted, a time window providingmuch opportunity for crustal evolution to takeplace. Understanding how oceanic crust evolvesis important from the perspectives of both basicscience (e.g., energy and mass exchange betweenthe Earth’s solid interior and the oceans) andsocietal impacts (e.g., subduction earthquake haz-ards). Our knowledge about this evolutionary pro-cess remains limited because of the inaccessibilityof the oceanic crust and the challenges associatedwith drilling and sampling it. For these reasons,geophysical surveying has played a major role inoceanic crustal studies during the past severaldecades.

[3] Early work on partitioning of the igneousoceanic crust into upper and lower seismic layers2 and 3 [e.g., Raitt, 1963] was almost entirelybased on interpretations of first arrival traveltimes.Researchers of the time noted that layer 2 velocitiesshowed significant variation from location to loca-tion and speculated that much of this variation mayoriginate within the top part of this layer. Basedmostly on its magnetic properties, layer 2 wasfurther subdivided into upper part A and lowerpart B [Talwani et al., 1971]. This subdivision oflayer 2 is still in use, and it is widely accepted thatthe steep vertical velocity gradient that defines theseismic layer 2A/2B boundary represents a poros-ity transition zone within the upper crust. However,the geologic nature of this porosity change contin-

ues to be debated, with two prevalent hypotheses:Layer 2A/2B boundary defines the geologicboundary between highly porous basaltic lavasand low-porosity diabase dykes [e.g., Herron,1982; Harding et al., 1993]; layer 2A/2B boundaryis an alteration front in the upper crust, probablywithin the extrusive section [e.g., Vera et al., 1990;Christeson et al., 2007]. These two hypothesesmay not be mutually exclusive for all mid-oceanridges. For regions with a steady state magmachamber, with little or no off-axis variation in layer2A thickness over time (e.g., East Pacific Rise), thewell-defined base of layer 2A is likely both alithologic boundary and alteration/permeabilityfront.

[4] The first to make the correlation between thechange in upper crustal seismic velocities andcrustal evolution were Houtz and Ewing [1976]based on an analysis of sonobouy data from theNorth Atlantic and Pacific. They concluded thatlayer 2A velocities increase from 2.8 to 3.3 km/s atthe ridge crests to >4.0 km/s on ridge flanks atabout 40 Ma-old crust, and speculated that layer2A likely changes with the passage of time throughinfilling of voids and cracks due to hydrothermalmineralization.

[5] Many approaches to investigating layer 2Avelocity have been applied since the study ofHoutz and Ewing [1976]. Most of these studieshave extracted upper crustal velocity informationfrom multichannel seismic (MCS) streamer datavia 1-D modeling and inversion techniques suchas interactive traveltime modeling [e.g., Vera andDiebold, 1994; Rohr, 1994], genetic algorithms[e.g., Hussenoeder et al., 2002], or waveforminversion [e.g., Collier and Singh, 1998]. Othertraveltime modeling studies were based on oceanbottom seismometer or expanding spread profiledata [e.g., Vera et al., 1990; Christeson et al., 1993;Grevemeyer et al., 1999]. The primary outcome of

GeochemistryGeophysicsGeosystems G3G3

nedimoviC et al.: evolution of uppermost oceanic crust 10.1029/2008GC002085

2 of 23

these seismic studies is twofold: (1) P wave veloc-ities within the layer 2A approximately double asit matures with increasing distance away fromthe ridge axis, which is the best documentedchange in the seismic structure of oceanic crustwith age [Purdy and Ewing, 1986]; (2) Layer 2Athickness (100–200 m) at fast spreading ridgesdoubles or triples within a few km of the ridgeaxis [e.g., Kent et al., 1994; Carbotte et al., 2000],while intermediate to slow spreading ridges aretypified by a thicker layer 2A section at the ridgecrest, exhibiting modest changes in thickness nearaxis [e.g., Blacic et al., 2004; Canales et al., 2005].These patterns of near-axis thickening may reflectdifferences in the accumulation of lavas linked toeruption parameters and ridge crest topography.Following the lead of Houtz and Ewing [1976],the velocity change within layer 2A has been com-monly attributed to precipitation of low-temperaturealteration minerals within the extrusive rocks asso-ciated with ridge flank hydrothermal circulation[Jacobson, 1992].

[6] Recent P wave velocity compilations indicatethat at a regional scale layer 2A doubles in velocitywithin � <10 Ma of crustal formation [Grevemeyerand Weigel, 1996; Carlson, 1998], much morequickly than originally interpreted by Houtz andEwing [1976]. This increase in seismic velocity oflayer 2A may not be a linear function of age. Amultistage evolution is suggested by studies on theflanks of the fast spreading East Pacific Rise[Grevemeyer and Weigel, 1997] with rapid velocityincrease at young ages (<1 Ma), a more gradualincrease up to 5 Ma, gentle increase up to �10 Ma,and no change at greater age. Grevemeyer andWeigel [1997] attribute the variable layer 2A hor-izontal velocity gradient, horizontal rate of changein layer 2A velocity with respect to ridge-normaldistance away from the spreading axis, to differentrates of crustal alteration associated with ridge axisand flank hydrothermalism. Numerical simulations[Fisher and Becker, 1995; Wang et al., 1997] andobservational studies [e.g., Langseth et al., 1988;Johnson et al., 1993] point to a close relationshipbetween hydrothermal upflow zones and basementrelief further suggesting locally variable, topogra-phy driven mineral precipitation within layer 2A,and therefore locally variable 2A velocity increase.

[7] Sediment blanketing, acting as a thermal insu-lator, has been proposed by Rohr [1994] to en-hance mineral precipitation within layer 2A andtherefore accelerate the velocity increase withcrustal aging. From analysis of a single MCS

profile crossing the eastern Endeavor flank of theJuan de Fuca ridge, Rohr [1994] found an abruptincrease in layer 2A velocity at very young crustalages (0.6–1.2 Ma) coincident with a transitionfrom sediment-free to fully sediment-buried oce-anic crust.

[8] Here we describe a systematic and uniformapproach to extracting upper crustal P wave veloc-ity and layer 2A thickness along hundreds of kilo-meters of long-streamer MCS profiles focused ongathering new information on oceanic crustal evo-lution. The primary motivation for this study wasto examine on a regional scale the role of basementage and sediment burial on crustal alteration due toridge flank hydrothermal circulation. We carriedout 1-D traveltime modeling of upper crustalstructure along ridge-normal MCS transects cross-ing the Endeavor, Northern Symmetric, and Cleftsegments of the Juan de Fuca ridge. Data from along hydrophone streamer are used to form a densegrid of analysis points, common midpoint (CMP)supergathers, spaced about every 3 km and extend-ing across transects about 300 km long. Thisdetailed approach to extracting upper crustal structureacross the ridge axis, extending about 150 km awayfrom the axis on both ridge flanks to 5–9 Ma-oldcrust, allows us to investigate both regional andlocal aspects of layer 2A evolution. The layer 2Avelocity and thickness study presented in this paperis supported by coincident seismic reflectionimages, which were formed first to help guidethe traveltime modeling. Along the Endeavor tran-sect, which was positioned to coincide with theODP/IODP Flank Flux experiment [Davis et al.,1992, 1997], drill hole data provide direct con-straints on crustal alteration and basement hydro-geologic conditions that can be used for correlationwith the upper crustal seismic properties.

2. Study Area and Seismic Data

[9] The Juan de Fuca ridge, located offshore west-ern North America (Figure 1), is the boundarybetween the Pacific and Juan de Fuca plates. This480 km long, NNE-oriented intermediate-ratespreading center (56 mm/a full spreading rate[e.g., Wilson, 1993]) comprises seven 50–100 kmlong segments, each with a distinct axial morphol-ogy and separated by nontransform offsets up to30 km in length. The Blanco and Sovanco fracturezones bound the Juan de Fuca ridge to the southand north, respectively. Several hundred kilometersto the east is the Cascadia subduction zone.


nedimoviC et al.: evolution of uppermost oceanic crust 10.1029/2008GC002085nedimoviC et al.: evolution of uppermost oceanic crust 10.1029/2008GC002085

3 of 23

[10] The western and eastern Juan de Fuca ridgeflanks are both crossed by propagator wakes butotherwise show prominent differences. Seamountsare found primarily on the Pacific plate, both asisolated edifices and in chains, several of which lieclose to and intersect the Juan de Fuca ridge axis[Davis and Karsten, 1986]. Sediments covering theeastern Juan de Fuca ridge flank are up to a fewkilometers thick at the northern Cascadia subduc-tion deformation front [e.g., Nedimovic et al.,

2003], and thin toward the ridge axis and south-ward away from the dominant source of terrige-nous sediment. The western Juan de Fuca ridgeflank is only sparsely sedimented, although sedi-ment cover generally increases to the north.

[11] In 2002, we carried out an extensive MCSsurvey of the Juan de Fuca ridge and its flanksduring R/V Maurice Ewing expedition EW0207.The MCS data were collected using a 6 km long,480 channel Syntron digital towed hydrophonearray, or streamer, with receiver groups spacedat 12.5 m. Streamer depth and feathering weremonitored with a mix of 13 depth-controlling and11 compass-enhanced DigiCourse birds, plus aGPS receiver on the tail buoy. A 10-element,49.2 L (3005 in3) tuned air gun array was usedas the source of acoustic energy, with shots fired ata 37.5 m spacing under GPS control. Postshotlistening time was 10.24 s and the returning acous-tic energy was sampled at a 2 ms rate. Data wererecorded on 3490E tapes in SEGD format using theSyntron Syntrack 480 seismic data acquisitionsystem. The recorded signal has a bandwidthranging from �2 Hz to over 100 Hz. The nominalCMP bin spacing is 6.25 m and the nominal datatrace fold is 80.

[12] Data were collected within the near axisregion [Canales et al., 2005, 2006; Carbotte etal., 2006; Van Ark et al., 2007] as well as alongthree ridge flank transects crossing the Cleft,Northern Symmetric and Endeavor segments andextending to crustal ages of 5–9 Ma [Nedimovic etal., 2005; Carbotte et al., 2008] (Figure 1). Cleft,the southernmost segment crossed by the longtransects, has a shallow and broad axial highnotched by a 2–3 km wide axial rift floodedwith recent lavas. Northern Symmetric (or Cobb)segment has a narrow, 1–2 km wide depressionbisecting the crest of a narrow and deeper axialhigh. At Endeavor, the northernmost segmentcrossed by our long transects, abundant faultingis observed in the floor of a 2–3 km wide axialtrough and there is little evidence for recenteruptions.

3. Data Analysis

3.1. Seismic Imaging

[13] The prestack processing strategy adopted forthe EW0207 MCS data consisted of standardstraight-line CMP bin geometry; F-K and bandpass(2-7-100-125 Hz) filtering to remove low-frequencytowing noise; amplitude correction for geometrical

Figure 1. The 2002 Juan de Fuca ridge flank seismicprofiles are plotted in orange over a Sun-illuminatedgray bathymetric map. Every 5000th common midpoint(CMP) is annotated along each profile. Thick blue lineunderlining a part of the Endeavor transect shows thelocation of earlier seismic and heat-flow studies. Redhexagons are drill hole locations. Thick black lines arethe interpreted traces of the ridge axis. Thin purple linesare magnetic isochrones [Wilson, 1988]. Their age isshown in the legend. Colored overlays outline areas ofthe crust formed during the normal magnetic periods.Gray overlay outlines the location of propagator wakes.The inset shows the location of the study area withrespect to North America.



4 of 23

spreading; surface-consistent minimum phase pre-dictive deconvolution to balance the spectrum andremove short period multiples; surface-consistentamplitude correction to balance anomalous shotand receiver-group amplitudes not related to wavepropagation; trace editing; velocity analysis usingthe velocity spectrum method; normal moveout(NMO) and dip moveout (DMO) corrections toalign signal for stacking; and CMP mute to removeoverly stretched data. The prepared prestack data,with and without automatic gain control, werethen stacked (averaged). The poststack processingincluded seafloor mute, primary multiple mute toreduce migration noise, trapezoidal bandpass fil-tering (2–7–100–125 Hz), and time migration tocollapse diffractions and reposition dipping reflec-tion events. To improve imaging within the oceanicplate below layer 2A/2B boundary, the late trav-eltime data were additionally bandpass filtered at2–7–20–40 Hz and mildly coherency filtered.

[14] Extracting an image of the layer 2A/2Bboundary, often referred to as the 2A event,requires a somewhat different processing schemebecause this event is not a true reflection [Hardinget al., 1993]. The prestack data preparation isidentical up to the velocity analysis, which is doneon bandpass filtered (2-7-40-60 Hz) constant ve-locity stacks. When the normal moveout velocitiesthat best flatten the retrograde branch of the 2Arefraction are chosen, the data are subsequentlystacked. The stacked layer 2A event is timemigrated and coherency filtered. A surgical mute,used to zero unwanted parts of seismic data, is thenapplied to extract only the layer 2A event. Mergingthe extracted layer 2A event with the reflectionsection forms the final, composite seismic image.

[15] An example of a composite seismic image,transect 17–3–1 (see Figure 1 for location), isshown in Figure 2a. Reflections from the seafloor,sediment interfaces, top of the igneous basement,top of the axial magma chamber, and layer 2A/2Bboundary event can all be identified. An enlargedsection from the eastern part of this transect isshown in Figure 2b to emphasize the high qualityof the collected data. Moho reflections, not shownin Figure 2, are also well imaged along all transectsand are the focus of previous studies [Nedimovic etal., 2005; Carbotte et al., 2008].

3.2. Traveltime Modeling

[16] We constructed about 6000 constant offsetstack CMP supergathers as potential input fortraveltime modeling of seismic arrivals. Each of

the CMP supergathers is formed by combining datatraces from 12 adjacent CMPs and then by stackingthe traces with identical nominal source-receiveroffsets. Some 300 of the CMP supergathers bestsuited for the analysis were then used for traveltimemodeling in 1-D.

[17] Because the field data are characterized by37.5 m shot spacing and 12.5 m receiver spacing,and the chosen CMP spacing is 6.25 m, the formedCMP supergathers have 480 data traces with stackfold of 2. We experimented with various CMPsupergather configurations, starting from combin-ing 6 adjacent CMPs and ending with combining24 with an increment of 6, thus forming super-gathers with 480 data traces and varying stack foldfrom 1 to 4. In most cases, the highest signal-to-noise ratio for the seismic arrivals of interest wasachieved when combining 12 adjacent CMP gath-ers. This indicates that for our profiles, the lateralvariations in two-way traveltime to igneous base-ment for traces with identical source-receiver offsetbecome large enough to negatively affect thesignal-to-noise ratio when stacking CMPs that aremore than �75 m apart.

[18] Our prestack data preparation for traveltimemodeling of seismic arrivals is identical to that forthe reflection imaging up to the NMO removal.CMP gathers without NMO correction are stackedinto CMP supergathers and linearly moved outusing velocity of 5500 m/s, the approximate ve-locity of the layer 2B refractions (Figures 3–5).CMP supergathers are then read into the JDseissoftware, which allows us to model reflection andrefraction traveltime arrivals for constant velocityand linear velocity gradient layers (see Appendix Afor a detailed description of JDseis software).

[19] Selecting and analyzing 300 CMP super-gathers spaced at about every 3 km was a signif-icant effort that resulted in a dense grid of 1-Dupper crustal velocity functions and layer 2Athicknesses. The selected 300 CMP supergathersare located over the smoothest sections of theigneous basement and have the highest signal-to-noise ratio. These gathers are characterized byprominent seismic arrivals of interest, includingtriplications caused by the high vertical velocitygradient in the lower part of layer 2A. This isimportant for our study because observations ofprominent layer 2A arrivals are not common; thesewaves are obscured by other stronger arrivals suchas reflections and diffractions. The CMP geometryalso greatly reduces possible effects of interface dipon the velocity analysis [Diebold and Stoffa, 1981].



5 of 23

[20] Three example CMP supergathers from eachof the investigated long transects crossing Endeav-or, Northern Symmetric, and Cleft ridge segments,with and without modeled seismic traveltime arriv-als, are shown in Figures 3, 4, and 5, respectively.Also shown are velocity models used to computethe seismic traveltime arrivals. For all transectsinvestigated, the three CMP supergathers presentedcover a range of geologic environments, from the

heavily sedimented eastern ridge flank, to thethinly sedimented or sediment-barren igneous cruston the western ridge flank or near the ridge axis.The traveltime modeling on CMP supergathers wascarried out by assuming a four or five layer model,depending on the presence or absence of sedimentsat the investigated location. Seawater and sedimentcolumn were modeled as constant velocity layers.Upper layer 2A, lower layer 2A, and uppermost

Figure 2. Seismic reflection image of the Endeavor transect 17–3–1 formed by analyzing MCS data collectedduring the 2002 EW0207 cruise. (a) Sediments, crystalline basement, layer 2A/2B boundary and axial magmachamber are generally all well imaged. (b) Strength of the layer 2A/2B boundary ‘‘reflection’’ can better bevisualized. Figure 2b is an enlarged image of the area outlined by the dashed box in Figure 2a. Where imaged, the 2A/2B seismic event varies in strength from strong to weak.



6 of 23

layer 2B were modeled as linear-vertical-gradientvelocity layers.

4. Results

[21] Seismic imaging and traveltime modelingresults for the three long transects crossing

Endeavor, Northern Symmetric, and Cleft ridgesegments are summarized in Figures 6 and 7. InFigure 6, the relationship between the imaged uppercrustal structure and layer 2A velocities is com-pared. Figure 7 shows the relationship between theimaged upper crustal structure and layer 2A thick-ness. The upper crustal structure in both Figures 6

Figure 3. Traveltime curves and the resulting model velocities for selected CMP supergathers (a) 1875, (b) 17,715,and (c) 22,085 from the Endeavor transect 17–3–1. The left images show the CMP supergathers corrected using alinear moveout (LMO) with velocity of 5500 m/s. Middle images show the same information as the correspondingimages in the left column but also include modeled traveltime arrivals (yellow lines) for the seafloor reflection,igneous basement reflection, layer 2A refraction (turning ray), and layer 2B refraction. Right images show thevelocity models that correspond to the fitted traveltime curves shown in the middle column images.



7 of 23

and 7 (top) is presented by traveltime picks ofseismic arrivals most important for this study, thosecoming from the seafloor, top of the igneous base-ment, layer 2A/2B boundary, and top of the axialmagma chamber. Seafloor and igneous basement areimaged continuously. The layer 2A/2B pseudo re-flection event is imaged along much of each tran-

sect. Average upper layer 2A velocities and averagewhole layer 2A velocities (Figure 6, middle andbottom), as well as upper layer 2A thicknesses andwhole layer 2A thicknesses (Figure 7, middle andbottom) are all presented in the context of distancefrom the ridge axis, crustal age, sediment cover, andpropagator wake distribution.

Figure 4. Traveltime curves and the resulting model velocities for selected CMP supergathers (a) 2460, (b) 14,300,and (c) 22,560 from the Northern Symmetric transect 34–32. Left images show the CMP supergathers correctedusing an LMO and velocity of 5500 m/s. Middle images show the same information as the corresponding images inthe left column but also include modeled traveltime arrivals (yellow lines) for the seafloor reflection, igneousbasement reflection, layer 2A refraction (turning ray), and layer 2B refraction. Right images show the velocity modelsthat correspond to the fitted traveltime curves shown in the middle column images.



8 of 23

[22] Uncertainties for measurements of seismic lay-er velocity and thickness presented in Figures 6and 7 vary at each CMP supergather locationdepending on many parameters such as datasignal-to-noise ratio, range of offsets over whichseismic arrivals can be identified, smoothness anddip of the layer boundaries, and sediment thick-

ness. We carried out a sensitivity analysis todetermine the range of permissible model parame-ters that fit the data. From the maximum andminimum fits, we estimated average uncertaintiesin individual CMP supergather measurements ofvelocity and thickness to be: Upper layer 2Aaverage velocity, ±150 m/s; lower layer 2A average

Figure 5. Traveltime curves and the resulting model velocities for selected CMP supergathers (a) 39,750,(b) 29,090, and (c) 12,460 from the Cleft 87–89–73–89a transect. Left images show the CMP supergatherscorrected using an LMO and velocity of 5500 m/s. Middle images show the same information as the correspondingimages in the left column but also include modeled traveltime arrivals (yellow lines) for the seafloor reflection,igneous basement reflection, layer 2A refraction (turning ray), and layer 2B refraction. Right images show thevelocity models that correspond to the fitted traveltime curves shown in the middle column images.



9 of 23

velocity, ±120 m/s; uppermost layer 2B averagevelocity, ±100 m/s; upper layer 2A thickness,±30 m; lower layer 2A thickness, ±10 m.

4.1. Seismic Images

[23] Reflection sections (Figures 2, 6, and 7) showthat the western and eastern Juan de Fuca ridgeflanks are evolving in a markedly different way dueto distinct sedimentary and volcanic histories.Hemipelagic sediments that thin southward aremuch thicker and more extensive on the easternridge flank. Enhanced accumulation of sediment onthe eastern flank is in large part caused by the

Figure 6. Crustal seismic structure and layer 2Avelocity results for the (a) Endeavor (17–3–1),(b) Northern Symmetric (34–32), and (c) Cleft (87–89–73–89a) MCS transects. Top sections show seismicstructure. Middle and lower parts show 2A velocityresults. Black stars connected with a thin black line areaverage upper 2A and whole 2A velocities from 1-Dtraveltime velocity analysis on CMP supergathers.Shaded areas outline the location of propagator wakes.Brown circles in Figure 6a are drill hole locations. Alldrill holes are missing 10 in front of the number (e.g., 23is drill hole 1023). CMP number, crustal age, and distancefrom the ridge axis are all given on the horizontal axis. Figure 7. Crustal seismic structure and layer 2A

thickness results for the (a) Endeavor (17–3–1),(b) Northern Symmetric (34–32), and (c) Cleft (87–89–73–89a) transects. Top sections show seismic structure.Middle and lower parts show 2A thickness. All theannotation in Figure 7 is identical to the annotationpresented in Figure 6.



10 of 23

morphology of the Juan de Fuca ridge, with itscooling and subsiding flanks forming basin-likedepositional environments and its elevated axialregion acting as a barrier that inhibits the transportof terrigenous sediment to the western flank.

[24] Sediments across the north-central section ofthe eastern ridge flank completely cover the igne-ous basement starting 20–30 km east from theridge axis (Figures 6a and 6b, top). However, thisthermal blanketing of igneous basement [e.g.,Wheat and Mottl, 1994] does not appear to bepresent in the southern section (Figure 6c, top)because the sedimentary cover in the area ispierced by many basement highs that may act asbasement ventilators. Sedimentary cover on thewestern flank is generally thin and discontinuousdue to the abundance of seamount volcanism[Davis and Karsten, 1986], although some isolatedpockets of thicker accumulations are identified inthe north.

4.2. Layer 2A Velocities and Thicknesses

[25] Layer 2A velocities (Figure 6) show a system-atic increase with distance from the ridge axis thatdiffers in magnitude between the seismic transectsand within each transect across the eastern andwestern ridge flanks. On the eastern flank, there isa significant increase in layer 2A velocities ofabout 60% along the Endeavor (�2.5–4.0 km/s)and Northern Symmetric (�2.3–3.7 km/s) trans-ects as the crust ages from 0 to 5–7 Ma. East ofthe Cleft ridge segment, along the southerntransect, the velocity increase is smaller (approx-imately 32%; �2.5–3.3 km/s). Similar to the eastCleft transect, more modest increase in layer 2Avelocity with crustal age is identified across thewestern ridge flank. For west Endeavor, NorthernSymmetric, and Cleft transects the velocity

increases are about 24% (�2.5–3.1 km/s), 30%(�2.3–3.0 km/s), and 24% (�2.5–3.1 km/s),over approximately 5, 4, and 9 Ma, respectively.

[26] There are no systematic variations in layer 2Athickness with distance from the ridge axis alongany of the transects (Figure 7). However, averagelayer 2A thicknesses along the eastern Endeavorand eastern Northern Symmetric flanks are lessthan that along the eastern Cleft and all of thewestern flanks (see Table 1). Similar transect-to-transect differences in average thicknesses areobserved for the low velocity gradient upper sec-tion of layer 2A but not for the high-velocitygradient lower section of layer 2A, whose thick-ness shows less variation (Table 1). Therefore,much of the variability in layer 2A thicknessmeasured from one ridge segment to another orig-inates within the low-velocity gradient upper sec-tion of layer 2A. Thickness estimates for veryyoung crust (<1 Ma), where constructional volca-nism may still be taking place, were excluded fromthe computation of average thicknesses shown inTable 1.

[27] The traveltime modeling results accuratelyreveal regional layer 2A velocity increases and2A thickness changes. Variability of layer 2Avelocity and thickness, from one 1-D analysis toanother, diminishes from north (Endeavor transect)to south (Cleft segment). Shorter wavelength var-iations in layer 2A velocity and thickness, super-imposed on the long-term systematic trends, arealso apparent. Nevertheless, because of the stilllimited lateral resolution of our dense 1-D study,we restrict our discussion to velocity and thick-ness anomalies that laterally extend for more than5–10 km.

5. Discussion

[28] The results presented in this work are com-pared to observations of layer 2A evolution evidentin existing global syntheses of Carlson [1998] andGrevemeyer et al. [1999]. Regionally, we compareour east Endeavor segment results with those fromthe coincident study of Rohr [1994] and the resultsof ODP/IODP Flank Flux experiment [Davis et al.,1992]. Additional constraints on the oceanic crustalstructure on the flanks of the Juan de Fuca ridge areavailable from a number of refraction and otherreflection studies conducted since the 1980s. How-ever, results from these studies (e.g., Cudrak andClowes [1993], Barclay and Wilcock [2004],McClymont and Clowes [2005], and Van Ark etal. [2007] for Endeavor; McClain and Lewis

Table 1. Computed Average Upper, Lower, and WholeLayer 2A Thicknesses for the Investigated Juan de FucaRidge Flanks

Ridge Flank

Average Layer 2A Thicknessa

Upper Lower Whole

East Endeavor 220 ± 70 100 ± 30 320 ± 80East North Symmetric 220 ± 70 110 ± 30 330 ± 80East Cleft 350 ± 90 110 ± 10 460 ± 90West Endeavor 350 ± 70 110 ± 20 460 ± 80West North Symmetric 340 ± 80 100 ± 10 440 ± 90West Cleft 270 ± 50 110 ± 10 380 ± 50

aAverage layer 2A thickness is measured in meters.



11 of 23

[1982] and Christeson et al. [1993] for NorthernSymmetric; and McDonald et al. [1994] andCanales et al. [2005] for Cleft) are not suitablefor a systematic regional-scale comparison with theresults extracted here as they either have limitedresolution within the shallowest crust or are toowidely spaced even when combined.

5.1. Impact of Crustal Age

5.1.1. Crustal Age and Layer 2A Velocity

[29] Average layer 2A P wave velocities across theJuan de Fuca ridge flanks systematically increasewith distance from the ridge axis, or with crustalage, as observed on ridge flanks globally [e.g.,Carlson, 1998; Grevemeyer and Bartetzko, 2004].However, this velocity increase is not uniform. Toquantify the rate of change of velocity with crustalage and test the hypothesis that layer 2A velocitiesdouble within 10 Ma [e.g., Purdy, 1987; Rohr,1994; Grevemeyer and Weigel, 1996], we fit thevelocity data from each ridge flank separately.Although the ‘‘eyeball’’ fit of the median velocitiesfrom Carlson [1998] global synthesis suggests apower law for the relationship between layer 2Avelocity and crustal age, we choose simple linearregression (least square fit of a straight line)because it results in smaller residuals and appearsmost appropriate for our data set. During fitting weremoved the data outliers due to anomalous crust inpropagator wakes. Examples include the propaga-tor wake along the eastern Endeavor ridge flankand the crust west of CMP 8000 on the Clefttransect, where crustal age cannot be determinedaccurately due to two closely spaced propagatorwakes (Figures 1 and 6). Final regressions weredone on �95% of the original measurements.

Computed horizontal velocity gradients for allinvestigated flanks are shown in Table 2.

[30] We use a linear extrapolation of the trendscomputed on �0–6 Ma-old crust to estimate thecrustal age at which the layer 2A velocity will bedouble that found on the ridge axis. Our resultssuggest that the layer 2A velocities across theeastern ridge flank for the Endeavor and NorthernSymmetric transects double by an age of �9 and�11 Ma, respectively, compared with �25, �20,and �23 Ma for the western Endeavor, NorthernSymmetric, and Cleft ridge flanks. For the layer 2Avelocities to double along the eastern Cleft ridgeflank it is estimated that the crust has to age for�21 Ma, significantly more than that along theeastern Endeavor and eastern Northern Symmetricridge flanks, and more in line with the estimatesfor the western ridge flanks. If the layer 2Avelocity increase is better described by a powerlaw [Carlson, 2004] or by a smoothly varyingpolynomial [Grevemeyer and Weigel, 1996], assuggested for the global velocity data set, thendoubling of layer 2A velocities along the Juan deFuca ridge flanks would take longer than estimatedhere.

[31] Another approach to evaluating crustal evolu-tion is to estimate crustal age at which layer 2Areaches velocities typical for mature oceanic crust[e.g., Grevemeyer et al., 1999]. For the easternflanks of the Endeavor, Northern Symmetric, andCleft ridge segments we estimate using linearextrapolation that the velocity of �4.3 km/s isreached at �7, �9, and �15 Ma, respectively.For the western flanks, the corresponding agesare �17, �17, and �16 Ma.

[32] Crustal ages computed by linear extrapolationof layer 2A velocities in both cases lead to similarconclusions. The results for the eastern Endeavorand eastern Northern Symmetric ridge flanks,which are blanketed with a sealing sedimentarycover, are consistent with the hypotheses that layer2A velocities double or reach mature oceanic crustvalues of �4.3 km/s within <10 Ma [Grevemeyeret al., 1999; Purdy, 1987; Grevemeyer and Weigel,1996]. The other four flanks, those with thin anddiscontinuous sediment cover, show a more grad-ual increase in layer 2A velocities, reaching�4.3 km/s within 15–17 Ma.

[33] To further investigate the relationship betweenthe Juan de Fuca ridge flank layer 2A velocitiesanalyzed here and the existing global database[Carlson, 1998; Grevemeyer et al., 1999] we plotthe two together in Figure 8. Velocities from our

Table 2. Computed Layer 2A P Wave Velocity andThickness Gradients for the Investigated Juan de FucaRidge Flanksa

Ridge Segment Eastern Flank Western Flank

Layer 2A Thickness GradientEndeavor �13 ± 14 �19 ± 13Northern Symmetric 15 ± 9 29 ± 25Cleft �8 ± 8 5 ± 3

Layer 2A P Wave Velocity GradientEndeavor 0.271 ± 0.013 0.103 ± 0.011Northern Symmetric 0.213 ± 0.008 0.117 ± 0.011Cleft 0.120 ± 0.005 0.109 ± 0.007

aLayer 2A thickness gradient is measured in m/Ma and layer 2A

P wave velocity gradient is measured in km s�1/Ma.



12 of 23

study show a similar trend of velocity increase withgreater crustal age but are less scattered than theglobal database. Several factors may contribute tothe lower scatter we observe. Unlike the globaldatabase, which includes studies of oceanic crustcreated at different spreading rates using a broad

range of seismic data sets (mostly wide-angle) andapplying different traveltime analysis techniques,the current study was of a relatively uniform inter-mediate-spread crust using a single MCS data setand a uniform traveltime analysis technique. Thenewly computed velocities also show a more grad-ual increase with crustal age than the prior compi-lations and do not seem to follow the power lawfunction as suggested by the mean velocities com-puted here for the global compilation of Carlson[1998] and Grevemeyer et al. [1999].

5.1.2. Crustal Age and Layer 2A Thickness

[34] Many research projects have been directedtoward investigating layer 2A [e.g., Kennett andOrcutt, 1976; Whitmarsh, 1978; Stephen andHarding, 1983; McClain and Atallah, 1985;Minshull et al., 1991; Christeson et al., 1994] sincethe seminal paper by Houtz and Ewing [1976].Because these investigations were usually doneover a relatively small section of the oceanic crust,they could not individually address one of the keyhypotheses proposed by Houtz and Ewing [1976]that layer 2A ‘‘thins’’ with increasing crustal ageand eventually becomes seismically indistinguish-able from layer 2B. By the mid-1990s, however,there were enough observations from individualseismic refraction experiments to compile a data-base [Grevemeyer and Weigel, 1996], expand it toinclude drillhole results [Carlson, 1998], and sta-tistically analyze it. From his analysis Carlson[1998] concludes that layer 2A does not disappearwith increasing crustal age but persists as a regionof relatively lower seismic velocities capping theoceanic crust.

[35] Statistical analysis done by Carlson [1998]cannot be applied to our data as most of ourtraveltime modeling was done on young oceaniccrust (0–5 Ma) and at no place, even within theolder crust (5–9 Ma), do the layer 2A velocitiesreach those found in the mature oceanic crust(�4.3 km/s). Therefore, to further contribute tothe work of Houtz and Ewing [1976] and Carlson[1998], we take a direct approach and investigatethe fate of layer 2A along the Juan de Fuca ridgeflanks by fitting the layer 2A thickness data fromeach ridge flank both independently and jointly.Long-term thickness trends are calculated exclud-ing results from young crust (0–1 Ma) that may beaffected by accretionary processes at the ridge axis.

[36] The computed thickness gradients (see Table 2)for one half of the investigated flanks are negative(Endeavor east, Endeavor west, and Cleft east), and

Figure 8. Average layer 2A P wave velocities as afunction of crustal age for the first 15 Ma of crustalevolution. Stars are both velocities compiled by Carlson[1998] for years 1976 to 1997 and velocities fromGrevemeyer et al. [1999] from an investigation designedto study upper crustal aging along the East Pacific Riseat 14�S. Solid black circles with error bars are meanvelocities from Carlson [1998] for ages <1 Ma, �1 and�5 Ma, and >5 and <20 Ma. Thin black line is thepower law fit of the global data set (stars) from Carlson[2004]. Hexagons are velocities from this study. Fullysedimented Endeavor and Northern Symmetric easternflanks are shown in red; partially sedimented Clefteastern flank is shown in green; barren or thinly andsparsely sedimented western flanks are shown in blue.Note that the velocity analysis done for this study was asignificant effort that resulted in three to four times asmany layer 2A velocity data points as there are in theglobal compilation [Carlson, 1998; Grevemeyer et al.,1999], but that these velocities characterize a relativelysmall and unique region of the oceanic crust.



13 of 23

for the other half are positive (Northern Symmetriceast, Northern Symmetric west, and Cleft west).This points to lack of systematic variation of layer2A thickness with increasing crustal age for theJuan de Fuca region. Furthermore, the errors in thecomputed thickness gradients (Table 2) are ofabout the same magnitude as the gradients them-selves, so thinning of layer 2A with increasingcrustal age is not statistically significant even forindividual flank transects with negative thicknessgradients.

[37] Figure 9a, where the relationship betweenlayer 2A thickness and crustal age for all sixtransects is jointly examined, lends further supportfor the conclusion that the layer 2A is not graduallythinning and disappearing in the Juan de Fucaregion. The computed thickness gradient of –2 ±4 m/Ma indicates that there is no systematic changein layer 2A thickness of some 400 m (400 ± 100 mmean value) with changing crustal age throughoutthe region. In Figure 9b, we plot the same data setas in Figure 9a but include thickness values deter-mined for the first 1 Ma of oceanic crust. Despitewhat appear to be several outliers in the youngoceanic crust, regression values for the whole data

set differ little from those computed for crust olderthan 1 Ma.

5.2. Impact of Sediment Cover

5.2.1. Sediment Cover and Layer 2AVelocity

[38] Sedimentary cover is believed to exert impor-tant control on the process of hydrothermal depo-sition within the upper igneous crust [e.g.,Jacobson, 1992; Alt, 1995]. In particular whenthe sediments are sealed and are insulating layer2A from the seawater above, the resulting higherbasement temperatures hasten the alteration ofbasalts. At the Endeavor segment, basement tem-peratures increase from <10�C near the ridge axis,where there are no sediment deposits, to 40–50�Csome 20 km east from the onset of the continuoussedimentary cover [Davis et al., 1992; Wheat andMottl, 1994]. For the low temperature alterationregime (<150�C) [Hunter et al., 1999], a signifi-cant change such as this could speed up thehydrothermal deposition which could affect layer2A velocity to a degree that can be detected usingseismic techniques in addition to the crustal agingeffect. This is possible because large changes in

Figure 9. Layer 2A thickness in the Juan de Fuca region as a function of oceanic crustal age. (a) Mean thicknessand thickness gradient for crustal ages 1 Ma and older. (b) Mean thickness and thickness gradient for the wholedatabase, from 0 to �9 Ma old crust.



14 of 23

seismic velocity can be generated with small re-duction in total porosity by preferential sealing oflow aspect ratio (thin) cracks [Wilkens et al., 1991].

[39] The differences in the increase of layer 2Avelocity as a function of crustal age betweenindividual ridge flanks (see section 5.1.1) aredirectly correlatable with the distribution and char-acteristics of the sedimentary cover within the Juande Fuca region (Figure 6). The largest velocityincrease with age is found in the north central part

of the eastern Juan de Fuca ridge-flank region (eastEndeavor and east Northern Symmetric flanks),where the most continuous and thickest sedimen-tary cover is imaged (Figure 8). In this area, thereare only a few isolated basement outcrops [e.g.,Fisher et al., 2003]. The rate of change in layer 2Avelocity as a function of crustal age decreases asthe continuity and thickness of the sediments arereduced southward to the moderately sedimentedeast Cleft flank and to the sparsely sedimentedigneous crust on the western ridge flanks.

[40] Comparison between the influence of thesediment thickness and continuity of sedimentarycover shows that the latter appears to have a greatereffect on the layer 2A evolution. Sealing sedimen-tary cover appears to double the effect of crustalaging on layer 2A velocities, as evidenced at theeast Endeavor and east Northern Symmetric ridgeflanks. The east Cleft and west Endeavor flanks,where significant sediment accumulations are con-fined to minibasins between large basement out-crops, show similar horizontal gradients in layer2A velocity as the west Northern Symmetric andwest Cleft flanks, which are sparsely sedimented orsediment starved (Table 2).

5.2.2. Comparison With Earlier Results

[41] Overall, our results are consistent with therapid increase in 2A velocities with sealing sedi-ment cover at the east Endeavor ridge flank foundin the earlier study of Rohr [1994]. However,despite general agreement in the velocity trend,there are important differences between the twoinvestigations in the upper crustal velocitiesobtained. In Figure 10a the estimated layer 2Avelocities from both studies are presented for the50 km section east of the Endeavor ridge wherethe MCS profile analyzed by Rohr [1994] and theEW0207 transect 17–3–1 are coincident. Withinthe 19 km closest to the ridge axis, where there isno measurable sedimentary cover, average layer2A velocities from this study (stars in Figure 10a)are lower than Rohr’s [1994] layer 2A intervalvelocities (green circles) by some 0.5 km/s, onaverage. This difference becomes much greater, onthe order of �2 km/s, for the sedimented section(19–50 km).

[42] Interval velocities for the Rohr [1994] studyare estimated from stacking velocities using theDix [1955] method that applies well to a horizon-tally stratified earth with constant velocity layersbut not so well for layer 2A with its velocitygradients and velocity gradient reflections. On the

Figure 10. Comparison of layer 2A velocities fromtwo coincident surveys over the Endeavor ridgesegment. (a) Purple stars connected by a thin black lineare new layer 2A velocities extracted by 1-D traveltimemodeling on CMP supergathers from the 2002 MCSdata set. Red dots are new layer 2A interval velocitiesextracted from analyzed stacking velocities. Green dotsrepresent the same type of information as red dots butare based on an earlier study and data by Rohr [1994].(b) Same as Figure 10a but with velocities representedby green dots corrected as suggested by Rohr [1994].(c) Sediment thickness from this study and basementtemperature results from the heat flow studies by Daviset al. [1997].



15 of 23

basis of comparisons between the Dix solutionsand forward traveltime models for known velocitystructure Rohr [1994] concludes that the Dix ap-proximation has overestimated the interval velocityof layer 2A by �0.3 km/s for the nonsedimentedsection, and by as much as 1 km/s for the sedi-mented section of the east Endeavor profile. Weuse this analysis to correct and replot velocitiesfrom Rohr [1994] for a better comparison of layer2A velocities [Figure 10b]. We also compute layer2A interval velocities for our study using the sameapproach as Rohr [1994] and plot them (red circlesin Figures 10a and 10b) without the correctionestimated by Rohr [1994].

[43] Figure 10b shows that the corrected Rohr[1994] velocities for the unsedimented section ofthe profile and for the first 10 km of the sedimentedsection mostly agree with velocities from thisstudy, although they remain marginally higher inparticular for the 15 km closest to the ridge axis.For the profile section from 30 to 50 km, correctedRohr [1994] velocities remain significantly higherthan the velocities from this study. At a greaterdistance from the ridge axis, from 50 to �150 kmalong the profile (not shown in Figure 10), layer2A velocities in Rohr [1994] maintain an averagevalue of �5.5 km/s (�4.5 km/s after correction),while velocities in this study only gradually in-crease to �4 km/s.

[44] Interval velocities computed in this study fromstacking velocities (Figures 10a and 10b) for com-parison with the same from Rohr [1994] are for themost part higher than the average layer 2A veloc-ities computed from the results of our traveltimemodeling. The mean difference between the twoestimates is �0.35 km/s for the unsedimented and�0.25 km/s for the sedimented section of theprofile. Interestingly, this is in agreement with theRohr [1994] estimate for the unsedimented crust(�0.3 km/s), but much less than was suggested forthe sedimented crust (up to 1 km/s). As such,interval velocities from this study show improvedagreement with the corrected interval velocitiesfrom Rohr [1994] (Figure 10b). However, evenafter applying the velocity corrections, the keyfeatures of the velocity mismatch persist.

[45] We attribute the disagreement in the estimatedlayer 2A velocities between the two studies to thetraveltime analysis techniques applied and thedifferent vintage data used. The most importantdifference between the two spatially coincidentdata sets is that ours was collected some 15 yearslater with about twice as long streamer and digital

technology, providing longer source-receiver offsetinformation and ensuring improved data quality.Having larger offsets turned out to be important forimaging and traveltime modeling as the layer 2Arefraction for the Juan de Fuca region turns atdistances between about 2 and 4.5 km (Figures 3,4, and 5). Far offsets are particularly important forinvestigating the sedimented sections of the ridgeflanks as the basement in this area becomes deeperand layer 2A refractions occur at the higher end ofthe 2 to 4.5 km offset range (Figures 3, 4, and 5).

[46] Data analysis for the Rohr [1994] study musthave been challenging for the sedimented igneousbasement where only a fraction of the layer 2Arefraction was recorded due to the short maximumoffset of 3–3.6 km. Individual interval velocityestimates from Rohr [1994] for this deep section ofthe igneous basement show much greater variabil-ity than for the unsedimented basement furthersupporting this conclusion. The difference betweensuccessive layer 2A interval velocity measure-ments, spatially separated by just a few kilometers,increases with increasing depth to the basement(see Figure 4 in the work of Rohr [1994]) reaching>3 km/s. Therefore, it is not surprising that thevelocities for the two studies agree within errorbounds for the shallow unsedimented igneouscrust close to the ridge axis where the data usedby Rohr [1994] fully image layer 2A event, but donot agree for the deep sedimented section of theigneous basement where the older data lack the faroffsets needed to constrain the layer 2A velocitiesaccurately.

[47] The main difference in the data analysis ap-proach between the two studies is that Rohr [1994]applied a method that models constant velocitylayers and is suited for reflection arrivals, whilewe used a modeling technique that also allows forvertical gradient velocity layers and is applicablefor investigating both reflection and refractionarrivals. This is significant as the layer 2A eventis a wide-angle retrograde refraction that for ourbasement depths does not occur at offsets of <2 kmand will therefore stack well at a wide range ofNMO velocities. Since for thin layers such as 2Aeven small variations in the NMO velocity lead tolarge variations in the derived interval velocity, it ischallenging to constrain accurately layer 2A inter-val velocities based on this approach [e.g., Hardinget al., 1993]. Interval velocities derived fromstacking velocities for this study show less varia-tion than those from Rohr [1994] because we havebetter constraints from long offsets, as already



16 of 23

discussed, but also because initial traveltime mod-eling using vertical gradient velocity layers indi-cated the need to use consistently the minimumNMO velocity that results in a high-quality layer2A stack.

5.2.3. Sediment Cover and Layer 2AThickness

[48] Potential impact of the sealing sedimentarycover on layer 2A thickness is investigated usingthe same approach as taken in section 5.1.2 tostudy the relationship between layer 2A thicknessand crustal age. Layer 2A thickness estimates fromthe east Endeavor and east Northern Symmetricridge flanks, where a generally continuous sedi-ment cover is observed, are plotted in Figure 11and simple linear regression is applied. The com-puted thickness gradient of 0 ± 6 m/Ma indicateslack of any systematic trend in layer 2A thickness,which after sediment burial measures on averagesome 330 ± 80 m. However, layer 2A along thefully sedimented east Endeavor and east NorthernSymmetric ridge flanks is on average 100 mthinner than along the sparsely and/or discontinu-

ously sedimented conjugate west flanks and bothflanks of the Cleft transect (average thickness of430 ± 80 m).

[49] The change from a thicker to a thinner layer2A on the fully sedimented ridge flanks appears tooccur within small sections of the seismic transectsthat extend some 10–20 km and mark the onset ofsediment burial. This change in layer 2A thicknessis more pronounced on the east Endeavor (CMPs19000–21000 in Figure 7a) than on the eastNorthern Symmetric ridge flank (CMPs 21000–23000 in Figure 7b). No further change in layer 2Athickness is observed beyond the onset of fullsediment burial region.

[50] We speculate that the change to thinner layer2A along the sedimented eastern flanks may largelyreflect alteration of the lower part of layer 2Awith onset of a warmer hydrothermal regime linkedto sediment blanketing [Davis et al., 1992] withenhanced precipitation of alteration minerals[Hunter et al., 1999]. Reduction in bulk porositywithin layer 2A through infilling of small voidsand cracks with mineral precipitates would beexpected to have the strongest effect on the highvelocity gradient lower section of layer 2A, char-acterized by lower intrinsic porosities inheritedfrom crustal formation. The closing of thin crackswithin this transition zone results in velocity in-crease that makes the lower part of the highgradient section of layer 2A seismically indistin-guishable from layer 2B, essentially leading to areduction in the layer 2A thickness.

[51] The thinner layer 2A along the fully sedimentburied sections of the east Endeavor and eastNorthern Symmetric flanks may have developedonly recently and over a short time period. Sedi-mentation history in this region is not directlycoupled with crustal aging and full sediment burialat the east Endeavor and east Northern Symmetricflanks may be as recent as 0.1 Ma. Hence fullburial and therefore more vigorous hydrothermalregime might have affected much of the nowblanketed sections of the ridge flanks within a timeperiod that is much shorter than the age of thesediment covered crust, leading to a simultaneouschange in layer 2A thickness across a significantcrustal age range.

5.3. Impact of Propagator Wakes

[52] Local variations in layer 2A velocity andthickness are evident at more than half of thecrossed propagator wakes (Figures 6 and 7). The

Figure 11. Layer 2A thickness measurements alongthe sediment blanketed sections of the east Endeavorand east Northern Symmetric transects as a function ofoceanic crustal age. Mean thickness and thicknessgradient for the plotted data are also shown.



17 of 23

most prominent of these anomalies is found alongthe east Endeavor transect where a propagatorwake is crossed 40–45 km from the ridge axis.Where a local change in layer 2A velocity isidentified, it is characterized by increase in thevelocity of the high-gradient, lower-layer 2A (0.2–

0.5 km/s) and little change in the low-gradient,upper-layer 2A velocity. This leads us to suggestthat propagator wakes may represent regions ofchannelized fluid flow that can potentially have asignificant effect on alteration history and base-ment fluid-flow patterns not previously recognized.Alternatively, the anomalous layer 2A structurecould be inherited from the time of crustal accre-tion [Bazin et al., 2001].

5.4. Correlation With Drill Hole StudiesAlong the East Endeavor Corridor

[53] Our east flank Endeavor seismic profile iscoincident with the borehole transect of Leg 168of the Ocean Drilling Program (ODP), providing aunique opportunity to constrain inferences from theseismic data on alteration of the shallow oceaniccrust. Ten holes along a transect �120 km longwere drilled through the sediment column and intothe shallow basement (mostly <10 m), with sedi-ment pore water samples collected to the sediment-basement interface [Elderfield et al., 1999]. Storedsamples were later analyzed to examine the con-ditions of fluid-rock interactions in the low tem-perature (<150�C) hydrothermal regime that variesfrom one with open communication with the sea-water column to one with very limited communi-cation with the ocean [Davis et al., 1997]. Resultsof these analyses indicate that the seawater passingthrough the oceanic crust has reacted with base-ment rocks [e.g., Elderfield et al., 1999] andalteration minerals in shallow basement rocks in-dicate a general trend of increasing alteration withdistance across the eastern flank from the ridgeaxis, as well as with depth of sediment burial[Hunter et al., 1999]. These results confirm thatthe progressive increase in 2A velocities we ob-serve are linked to alteration of the uppermostcrust.

[54] In Figure 12, we compare the change inphysical properties of the uppermost crust, namelyP wave velocities from this study, with the changein the strontium isotope ratio and sulphate contentof the borehole basement fluids along the eastEndeavor ODP transect. On a regional scale, in-crease in the layer 2A P wave velocity correlateswell with major changes in these two pore fluidchemical parameters indicative of increased fluid-rock reactions. Other chemical parameters [seeElderfield et al., 1999], not shown in Figure 12,show similar correlation. The majority of geochem-ical change occurs within the first 30–40 km fromthe onset of the sedimentary cover, coincident with

Figure 12. Physical properties of the uppermostoceanic crust and selected chemical properties of thedrill hole basement fluids as a function of eastwarddistance from the Endeavor ridge crest: (a) Subset ofseismic velocities shown in Figure 6; (b) temperature;(c) strontium isotopes; (d) sulphate. Triangles indicateseawater (symbol positioned at distance of sedimentonlap onto ridge flank); circles indicate basal porewaters from upwelling sites 1030 and 1031; dotsindicate basal pore waters from other sites (joined bysolid line); squares indicate basement fluid from Site1026. Plots of chemical properties of basement fluidsmodified from Davis et al. [1997] and Elderfield et al.[1999]. Shaded area outlines the location of apropagator wake.



18 of 23

the prominent basement scarp �19 km eastwardfrom the ridge crest. Velocities do not start tochange immediately with the onset of sedimentarycover but rather some 5–10 km eastward from thislocation. Beyond this point velocities, when theeffect of the propagator wake is removed, appear toincrease approximately monotonously with dis-tance away from the ridge crest.

[55] Notably, the propagator wake is the site ofODP Leg 168 drill holes 1030, 1031, and 1028(Figures 6 and 7: 1030 and 1031 on the youngcrust side, 1028 across the age discontinuity onolder crust). Basalts sampled at the base of hole1025 to the west of the wake are vesicular massiveferrobasalts [Davis et al., 1997] and can now berecognized as the Fe-rich basalts often found inproximity to propagating ridge tips. Basementrecovery for holes drilled within the wake itselfwas negligible which could reflect more shearedand fractured basement associated with these struc-tures. The chemical signature of the fluids sampledover the propagator wake show either anomalousvalues or a major change in spatial gradient. Thisfurther supports our suggestion that propagatorsmay locally modify the regional basement fluidflow and could be characterized by a uniquealteration history.

6. Conclusions

[56] We described a systematic and uniform ap-proach to extracting upper crustal P wave velocityand layer 2A thickness from our 2002 Juan deFuca MCS data by 1-D traveltime modeling at�3 km intervals along 300 km long transectscrossing the Endeavor, Northern Symmetric, andCleft ridge segments. For this analysis we con-structed 6000 constant offset stack CMP super-gathers and selected 300 best suited for 1-Dtraveltime modeling. To support the traveltimemodeling, we first formed coincident seismicreflection images.

[57] Regionally, our results show a direct correla-tion between an increase in layer 2A velocity andincreasing crustal age. However, the identifiedvelocity increase varies across the investigatedridge flanks and supports a first-order effect of fullsediment burial on layer 2A evolution. For theflanks blanketed with a continuous sealing sedi-mentary cover, the velocity increases at about twicethe rate observed for ridge flanks with sparse ordiscontinuous sediment cover (0.24 versus 0.11 kms�1/Ma). Extrapolation of the rate of velocity

change with crustal age suggests that on the flankswith continuous sediment cover layer 2A velocitymay reach values typical of mature oceanic crust(�4.3 km/s) in <10 Ma. For the sparsely sedi-mented or sediment barren flanks velocities increasemore slowly (�16 Ma to reach �4.3 km/s).The correlation between layer 2A velocity andsediment cover is likely due to more rapid precip-itation of alteration minerals in the porous uppercrust as the hydrothermal regime evolves from onedominated by open exchange with the water col-umn to a regime that is effectively closed toseawater exchange by the sealing sedimentaryblanket.

[58] The computed ridge-normal thickness gra-dients show that layer 2A does not systematicallythin and disappear in the Juan de Fuca region withincreasing crustal age, although the increasingseismic velocities indicate progressive alteration.Layer 2A persists as a region of relatively lowerseismic velocities capping the oceanic crust regard-less of the presence and type or lack of thesedimentary cover. However, layer 2A along thefully sedimented ridge flanks is on average 100 mthinner than along the sparsely and discontinuouslysedimented flanks (330 ± 80 m versus 430 ± 80 m).The transition from thinner to thicker layer 2Aappears to take place in the region some 10–20 kmwide and centered on the onset of sediment burial.Change in layer 2A thickness beyond the onset offull sediment burial region is not observed.

[59] We speculate that the change to thinner layer2A along the sedimented eastern flanks is causedby alteration of the lower part of layer 2A withonset of a warmer hydrothermal regime linked tosediment blanketing with enhanced precipitation ofalteration minerals. Consequent closing of thincracks in the low-porosity lowermost layer 2Aresults in velocity increase that is significantenough to make this section of layer 2A seismi-cally indistinguishable from layer 2B, essentiallyreducing the layer 2A thickness. The thinner layer2A along the fully sediment buried sections of theeast Endeavor and east Northern Symmetric flanksmay have developed only recently and over a shorttime period. This is possible due to decouplingbetween the sedimentation history and crustal ag-ing that provides a mechanism for recent fullsediment blanketing and simultaneous layer 2Athinning within sedimented crust of different age.

[60] Locally, our results show correlation betweenthe location of propagator wakes and a change inlayer 2A thickness and velocity. This indicates that



19 of 23

either some propagator wakes represent zones ofenhanced fluid flow and enhanced precipitation ofalteration minerals, or extrusive sections in theseareas are formed in a unique way. The chemicalsignatures of the fluids sampled by ODP/IODPdrill holes along the east Endeavor transect supportthis observation. Other short wavelength variationsin 2A structure are evident in our analysis butdetermining the origin of these anomalies willrequire application of 2-D analysis techniques, inparticular 2-D tomography.

Appendix A: JDseis Software

[61] JDseis is a tool for display and interactiveanalysis of seismic data (Figure A1), includingstacked sections and t-p and t-x ensembles, wheret is intercept or vertical delay time, p is slowness, t

is two-way traveltime, and x is source-receiveroffset. One of the most important t-x data imagemanipulations is the use of a ‘‘reducing velocity’’(VREDUCTION) to apply a linear moveout (LMO) ora deck-of-cards linear time shift:

tREDUCED ¼ tRECORDED � x=VREDUCTION ðA1Þ

This time shift greatly improves the visualdiscrimination of various reflection and refractionbranches. During the ray tracing, predicted travel-time arrivals are superimposed upon these imagesand constantly redrawn as the model is developedand modified.

A1. JDseis Ray Trace Models

[62] In general, distinct, coherent body wave arriv-als in MCS data arise from one of three physicalprocesses: precritical reflection, postcritical reflec-

Figure A1. JDseis software main window with various pull down menus and a couple of utility windows open.



20 of 23

tion, and refraction. Two types of refracted arrivalshave been used for interpretation: head waves andturning, or diving rays. A head wave arises when awave intersects a velocity discontinuity at oneparticular angle (the ‘‘critical angle’’) at which itsrefraction creates a phase that travels along thediscontinuity. Virtually all of the incident energy,however, is reflected upward, the critical reflection.On the other hand, a wave traveling at an anglewithin a vertical velocity gradient will refractcontinuously, following a circular path. When thesewaves are turned enough that they begin to travelupward, they are in fact ‘‘totally refracted.’’

[63] A generalized layer represented by a verticalvelocity gradient and a velocity discontinuity at itsbottom can generate all four types of arrivals(Figure A2). JDseis models are made up of a stackof such layers, any of which may or may notinclude a vertical velocity gradient. Analysis iscarried out from top down, since the results ofeach overlying layer affect the raypaths in layersbelow. When creating a new model, layers areadded or inserted as needed, and their two waytimes and velocities altered to obtain the best fit‘‘by eye’’ to observed arrivals in the data. As thesechanges are made, a graphic display of the currentmodel is updated (Figure A1) so that the creationof erroneous and unrealistic layers can be avoided.

While modeling gathers along an MCS line, it ismost productive to modify existing models fromadjacent locations. In this way, horizons can betracked consistently along the line. Identification ofprimary arrivals in the data and choice of whicharrivals to fit is a crucial step in the analysis, andaccuracy depends heavily on the experience of andcare taken by the interpreter. Best results areusually attained through working iteratively withgroups of records acquired in geologically similarenvironments. The same arrivals should appear oneach record, but due to differences in water depthand other layer thicknesses, the exact position andappearance of various arrivals will differ, making iteasier to decide whether or not they in fact exist.

A2. Accuracy

[64] It is difficult to properly quantify the accuracyof any method of velocity analysis. This is due tothe great variety in the appearance of the arrivalsand in the range of offsets over which they may beobserved. The default velocity step used in theJDseis ray tracing is 5 m/s, and when reflected orrefracted arrivals can be observed over a significantenough range of offsets, the velocity resolutionunder ideal circumstances is within this range. Thiscondition is usually met with reflections withinhomogeneous layers. Postcritically refracted arriv-

Figure A2. Plot showing seismic arrivals modeled with JDseis. Shown is (top left) a two-layer model with (topright) the velocity structure. The model generates four types of seismic arrivals: precritical reflection, postcriticalreflection, head waves, and turning, or diving rays, showing (bottom) these arrivals in t-x and t-p domain.



21 of 23

als are usually seen within a smaller span ofsource-receiver offsets, and the velocity resolutionmay be as high as 10–25 m/s, depending on thevertical velocity gradients.

Acknowledgments

[65] This research was supported by National Science Foun-

dation grants OCE-00-02488, OCE-00-02551, and OCE-00-

02600. We are grateful to Captain Mark Landow, Science

Officer Joe Stennet, the crew, and the scientific and technical

party of the R/VMaurice Ewing Cruise 02–07 for their support

and help during the data acquisition. We thank R. L. Carlson,

G. L. Christeson, and E. E. Davis for their critical reviews.

References

Alt, J. C. (1995), Subseafloor processes in mid-ocean ridgehydrothermal systems, in Seafloor Hydrothermal Systems:Physical, Chemical, Biological, and Geological Interactions,Geophys. Monogr. Ser., vol. 91, edited by S. Humphris et al.,pp. 85–114, AGU, Washington, D.C.

Barclay, A. H., and W. S. D. Wilcock (2004), Upper crustalseismic velocity structure and microearthquake depths at theEndeavour Segment, Juan de Fuca Ridge, Geochem. Geo-phys. Geosyst., 5, Q01004, doi:10.1029/2003GC000604.

Bazin, S., et al. (2001), Three-dimensional shallow crustalemplacement at the 9�030N overlapping spreading centeron the East Pacific Rise: Correlations between magnetizationand tomographic images, J. Geophys. Res., 106, 16,101–16,117, doi:10.1029/2001JB000371.

Blacic, T., G. Ito, J. P. Canales, R. S. Detrick, and J. M. Sinton(2004), Constructing the crust along the Galapagos Spread-ing Center 91.3�–95.5�W: Correlation of seismic layer 2Awith axial magma lens and topographic characteristics,J. Geophys. Res., 109, B10310, doi:10.1029/2004JB003066.

Canales, J. P., R. S. Detrick, S. M. Carbotte, G. M. Kent, J. B.Diebold, A. J. Harding, J. M. Babcock, M. R. Nedimovic,and E. M. van Ark (2005), Upper crustal structure and axialtopography at intermediate-spreading ridges: Seismic con-straints from the southern Juan de Fuca Ridge, J. Geophys.Res., 110, B12104, doi:10.1029/2005JB003630.

Canales, J. P., S. C. Singh, R. S. Detrick, S. M. Carbotte, A. J.Harding, G. M. Kent, J. B. Diebold, J. Babcock, and M. R.Nedimovic (2006), Seismic evidence for variations in axialmagma chamber properties along the southern Juan de FucaRidge, Earth Planet. Sci. Lett., 246, 353–366, doi:10.1016/j.epsl.2006.04.032.

Carbotte, S. M., G. Ponce-Correa, and A. Solomon (2000),Evaluation of morphological indicators of magma supplyand segmentation from a seismic reflection study of theEPR 15�300–17�N, J. Geophys. Res., 105, 2737–2759,doi:10.1029/1999JB900245.

Carbotte, S. M., R. S. Detrick, A. J. Harding, J. P. Canales,J. Babcock, G. M. Kent, E. M. Van Ark, M. R. Nedimovic,and J. B. Diebold (2006), Rift topography linked to magma-tism at the intermediate spreading Juan de Fuca Ridge, Geol-ogy, 34, 209–212, doi:10.1130/G21969.1.

Carbotte, S. M., M. R. Nedimovic, J. P. Canales, G. M. Kent,and A. J. Harding (2008), Variable crustal structure along theJuan de Fuca Ridge: Influence of on-axis hotspots and abso-lute plate motions, Geochem. Geophys. Geosyst., 9, Q08001,doi:10.1029/2007GC001922.

Carlson, R. L. (1998), Seismic velocities of the uppermostigneous crust: Age dependence and the fate of layer 2A,J. Geophys. Res., 103, 7069–7077, doi:10.1029/97JB03577.

Carlson, R. L. (2004), Seismic properties of layer 2A at 11 Ma:Results of a vertical seismic profile at Ocean DrillingProgram Site 1243, Geophys. Res. Lett., 31, L17601,doi:10.1029/2004GL020598.

Christeson, G. L., G. M. Purdy, and K. M. M. Rohr (1993),Structure of the Northern Symmetrical Segment of the Juande Fuca Ridge, Mar. Geophys. Res., 15, 219–240,doi:10.1007/BF01204234.

Christeson, G. L., G. M. Purdy, and G. J. Fryer (1994), Seismicconstraints on shallow crustal emplacement processes at thefast spreading East Pacific Rise, J. Geophys. Res., 99,17,957–17,973, doi:10.1029/94JB01252.

Christeson, G. L., K. D. McIntosh, and J. A. Karson (2007),Inconsistent correlation of seismic layer 2a and lava layerthickness in oceanic crust, Nature, 445, 418 – 421,doi:10.1038/nature05517.

Collier, J. S., and S. C. Singh (1998), Poisson’s ratio structureof young oceanic crust, J. Geophys. Res., 103, 20,981–20,996, doi:10.1029/98JB01980.

Cudrak, C. F., and R. M. Clowes (1993), Crustal structure ofEndeavour ridge segment, Juan de Fuca ridge, from adetailed seismic refraction survey, J. Geophys. Res., 98,6329–6349, doi:10.1029/92JB02860.

Davis, E. E., and J. L. Karsten (1986), On the cause of theasymmetric distribution of seamounts about the Juan de FucaRidge; ridge-crest migration over a heterogeneous astheno-sphere, Earth Planet. Sci. Lett., 79, 385–396, doi:10.1016/0012-821X(86)90194-9.

Davis, E. E., et al. (1992), Flankflux, An experiment to studythe nature of hydrothermal circulation in young oceaniccrust, Can. J. Earth Sci., 29, 925–952.

Davis, E. E., A. T. Fisher, and J. V. Firth (1997), The Ship-board Scientific Party, Hydrothermal circulation in the ocea-nic crust: Eastern flank of the Juan de Fuca Ridge, Proc.Ocean Drill. Program Initial Rep., 168A, 1–470.

Diebold, J. B., and P. L. Stoffa (1981), The traveltime equa-tion, tau-p mapping and inversion of common midpoint data,Geophysics, 46, 238–254, doi:10.1190/1.1441196.

Dix, C. H. (1955), Seismic velocities from surface measure-ments, Geophysics, 20, 68–86, doi:10.1190/1.1438126.

Elderfield, H., C. G. Wheat, M. J. Mottl, C. Monnin, andB. Spiro (1999), Fluid and geochemical transport throughoceanic crust: A transect across the eastern flank of the Juande Fuca Ridge, Earth Planet. Sci. Lett., 172, 151–165,doi:10.1016/S0012-821X(99)00191-0.

Fisher, A. T., and K. Becker (1995), Correlation betweenseafloor heat flow and basement relief: Observational andnumerical examples and implications for upper crustalpermeability, J. Geophys. Res., 100, 12,641–12,657,doi:10.1029/95JB00315.

Fisher, A. T., et al. (2003), Hydrothermal recharge and dis-charge across 50 km guided by seamounts on a young ridgeflank, Nature, 421, 618–621, doi:10.1038/nature01352.

Grevemeyer, I., and A. Bartetzko (2004), Hydrothermal agingof ocenic crust; inferences from seismic refraction and bore-hole studies, in Hydrogeology of the Oceanic Lithosphere,edited by E. E. Davis and H. Elderfield, pp. 128–150,Cambridge Univ. Press, New York.

Grevemeyer, I., and W. Weigel (1996), Seismic velocities ofthe uppermost igneous crust versus age, Geophys. J. Int.,124, 631–635, doi:10.1111/j.1365-246X.1996.tb07041.x.

Grevemeyer, I., and W. Weigel (1997), Increase of seismicvelocities in upper oceanic crust: The ‘‘superfast’’ spreading



22 of 23

East Pacific Rise at 14�140S, Geophys. Res. Lett., 24, 217–220, doi:10.1029/96GL04005.

Grevemeyer, I., N. Kaul, H. Villinger, and W. Weigel (1999),Hydrothermal activity and the evolution of the seismic prop-erties of the upper oceanic crust, J. Geophys. Res., 104,5069–5079, doi:10.1029/1998JB900096.

Harding, A. J., G. M. Kent, and J. A. Orcutt (1993), A multi-channel eismic investigation of upper crustal structure at9�N on the East Pacific Rise: Implications for crustal accre-tion, J. Geophys. Res., 98, 13,925–13,944, doi:10.1029/93JB00886.

Herron, T. J. (1982), Lava flow layer-East Pacific Rise, Geo-phys. Res. Lett., 9, 17–20, doi:10.1029/GL009i001p00017.

Houtz, R., and J. Ewing (1976), Upper crustal structure as afunction of plate age, J. Geophys. Res., 81, 2490–2498,doi:10.1029/JB081i014p02490.

Hunter, A. G., P. D. Kempton, and P. Greenwood (1999), Low-temperature fluid-rock interaction-an isotopic and mineralo-gical perspective of upper crustal evolution, eastern flank ofthe Juan de Fuca Ridge (JdFR), ODP Leg 168, Chem. Geol.,155, 3–28, doi:10.1016/S0009-2541(98)00138-7.

Hussenoeder, S. A., R. S. Detrick, G. M. Kent, H. Schouten,and A. J. Harding (2002), Fine-scale seismic structure ofyoung upper crust at 17�200S on the fast spreading EastPacific Rise, J. Geophys. Res., 107(B8), 2158,doi:10.1029/2001JB001688.

Jacobson, R. S. (1992), Impact of crustal evolution on changesof the seismic properties of the uppermost ocean crust, Rev.Geophys., 30, 23–42, doi:10.1029/91RG02811.

Johnson, H. P., K. Becker, and R. Von Herzen (1993), Near-axisheat flow measurements on the northern Juan de Fuca Ridge:Implications for fluid circulation in oceanic crust, Geophys.Res. Lett., 20, 1875–1878, doi:10.1029/93GL00734.

Kennett, B. L. N., and J. A. Orcutt (1976), Comparison oftravel time inversions for marine refraction profiles, J. Geo-phys. Res., 81, 4061–4070, doi:10.1029/JB081i023p04061.

Kent, G. M., A. J. Harding, J. A. Orcutt, R. S. Detrick, J. C.Mutter, and P. Buhl (1994), Uniform accretion of oceaniccrust south of the Garrett transform at 14�150S on the EastPacific Rise, J. Geophys. Res., 99, 9097–9116, doi:10.1029/93JB02872.

Langseth, M. G., M. J. Mottl, M. A. Hobart, and A. Fisher(1988), The distribution of geothermal and geochemical gra-dients near Site 501/504: Implications for hydrothermal cir-culation in the oceanic crust, Proc. Ocean Drill. ProgramInitial Rep., 111, 23–32.

McClain, J. S., and C. A. Atallah (1985), The structure of youngoceanic crust bear a very fast spreading ridge, Geophys. Res.Lett., 12, 689–692, doi:10.1029/GL012i010p00689.

McClain, J. S., and B. T. R. Lewis (1982), Geophysical evidencefor the absence of a crustal magma chamber under the NorthernJuan de Fuca Ridge: A contrast with ROSE results, J. Geophys.Res., 87, 8477–8489, doi:10.1029/JB087iB10p08477.

McClymont, A. F., and R. M. Clowes (2005), Anomalouslithospheric structure of Northern Juan de Fuca plate-a con-sequence of oceanic rift propagation?, Tectonophysics, 406,213–231, doi:10.1016/j.tecto.2005.05.026.

McDonald, M. A., J. A. Hildebrand, and S. C. Webb (1994),Seismic structure and anisotropy of the Juan de Fuca Ridgeat 45�N, J. Geophys. Res., 99, 4857–4874, doi:10.1029/93JB02801.

Minshull, T. A., R. S. White, J. C. Mutter, P. Buhl, R. S.Detrick, C. A. Williams, and E. Morris (1991), Crustal struc-ture at the Blake Spur Fracture Zone from expanding spreadprofiles, J. Geophys. Res., 96, 9955–9984, doi:10.1029/91JB00431.

Nedimovic, M. R., R. D. Hyndman, K. Ramachandran, andG. D. Spence (2003), Reflection signature of seismic andaseismic slip on the northern Cascadia subduction interface,Nature, 424, 416–420, doi:10.1038/nature01840.

Nedimovic, M. R., S. M. Carbotte, A. J. Harding, R. S.Detrick, J. P. Canales, G. M. Kent, M. Tischer, J. B. Diebold,and J. M. Babcock (2005), Frozen subcrustal magma lenses,Nature, 436, 1149–1152, doi:10.1038/nature03944.

Purdy, G. M. (1987), New observations of the shallow seismicstructure of young oceanic crust, J. Geophys. Res., 92,9351–9362, doi:10.1029/JB092iB09p09351.

Purdy, G. M., and J. Ewing (1986), Seismic structure of theocean crust, in The Geology of North America, vol. M, TheWestern North Atlantic Region, edited by P. R. Vogt and B. E.Tucholke, pp. 313–370, Geol. Soc. of Am., Boulder, Colo.

Raitt, R. W. (1963), The crustal rocks, in The Sea, vol. 3, TheEarth Beneath the Sea History, edited by M. N. Hill, pp. 85–102, Wiley-Interscience, New York.

Rohr, K. M. M. (1994), Increase of seismic velocities in upperoceanic crust and hydrothermal circulation in the Juan deFuca plate, Geophys. Res. Lett., 21, 2163 – 2166,doi:10.1029/94GL01913.

Stephen, R. A., and A. J. Harding (1983), Travel time analysisof borehole seismic data, J. Geophys. Res., 88, 8289–8298,doi:10.1029/JB088iB10p08289.

Talwani, M., C. C. Windisch, and M. G. Langseth (1971),Reykjanes ridge crest: A detailed geophysical study, J. Geo-phys. Res., 76, 473–517, doi:10.1029/JB076i002p00473.

Van Ark, E. M., R. S. Detrick, J. P. Canales, S. Carbotte, A. J.Harding, G. M. Kent, M. R. Nedimovic, W. S. D. Wilcock,J. B. Diebold, and J. M. Babcock (2007), Seismic structureof the Endeavour segment Juan de Fuca Ridge: Correlationswith seismicity, faulting and hydrothermal activity, J. Geo-phys. Res., 112, B02401, doi:10.1029/2005JB004210.

Vera, E. E., and J. B. Diebold (1994), Seismic imaging ofoceanic layer 2A between 9�300N and 10�N on the EastPacific Rise from two-ship wide-aperture profiles, J. Geo-phys. Res., 99, 3031–3041, doi:10.1029/93JB02107.

Vera, E. E., J. C. Mutter, P. Buhl, J. A. Orcutt, A. J. Harding,M. E. Kappus, R. S. Detrick, and T. M. Brocher (1990), Thestructure of 0- to 0.2-m.y.-old oceanic crust at 9�N on theEast Pacific Rise from expanded spread profiles, J. Geophys.Res., 95, 15,529–15,556, doi:10.1029/JB095iB10p15529.

Wang, K., J. He, and E. E. Davis (1997), Influence of base-ment topography on hydrothermal circulation in sediment-buried igneous oceanic crust, Earth Planet. Sci. Lett., 146,151–164, doi:10.1016/S0012-821X(96)00213-0.

Wheat, G. C., and M. J. Mottl (1994), Hydrothermal circula-tion, Juan de Fuca Ridge eastern flank: Factors controllingbasement water composition, J. Geophys. Res., 99, 3067–3080, doi:10.1029/93JB01612.

Whitmarsh, R. B. (1978), Seismic refraction studies of upperigneous crust in North Atlantic and porosity estimates forlayer 2, Earth Planet. Sci. Lett., 37, 451–464, doi:10.1016/0012-821X(78)90061-4.

Wilkens, R. H., G. J. Fryer, and J. Karsten (1991), Evolution ofporosity and seismic structure of upper oceanic crust: Impor-tance of aspect ratios, J. Geophys. Res., 96, 17,981–17,995,doi:10.1029/91JB01454.

Wilson, D. S. (1988), Tectonic history of the Juan de Fucaridge over the last 40 million years, J. Geophys. Res., 93,11,863–11,876, doi:10.1029/JB093iB10p11863.

Wilson, D. S. (1993), Confidence intervals for motion anddeformation of the Juan de Fuca plate, J. Geophys. Res.,98, 16,053–16,071, doi:10.1029/93JB01227.



23 of 23

Upper crustal evolution across the Juan de Fuca ridge flanks · 4] The first to make the...

Documents

Transcript of Upper crustal evolution across the Juan de Fuca ridge flanks · 4] The first to make the...